Copyright c 1997 Charter Media, Inc. All rights reserved.
The purpose of StreetBeat is to provide you with additional insights on the
market from recognized financial experts on (and off) Wall Street. It should
be noted that the views and opinions expressed by the panelists below are
not necessarily those of Briefing.com.
This week's topic: Semiconductor Equipment Industry
----------------------------------------------------------------------------
----
Panelists
a.. Gerald Fleming, Director of Research and Semiconductor Equipment Analyst
at Sutro & Co.
ÿÿÿ a.. Jeff Middleswart, Research Analyst at David W. Tice & Associates.
ÿÿÿ a.. Min Pang, Semiconductor Equipment Analyst at Cowen & Co.
ÿÿÿ Q&ntctA
Briefing: To what extent will the slowdown in SE Asia affect those
semiconductor equipment companies that have significant exposure to the
region? Will a boom in SE Asian exports resulting from currency devaluations
help to support demand for imports of semiconductor equipment?
Jerry Fleming: SE Asia is of major importance to the semiconductor industry.
However, currency devaluations do not have a direct impact because
semiconductor equipment is a dollar-dominated industry. The significant
factor is the tightening in lending standards by SE Asian banks. For
example, Korea, which represents about 15% of the semiconductor equipment
market, has seen the won depreciate considerably. Reports are that the big
three Korean DRAM manufacturers, Samsung, Hyundai, and LG (Lucky Goldstar),
are delaying plans for expansion and have been dumping DRAMs on the market
in an effort to raise cash since bank loans are restricted. Likewise, three
Taiwanese joint ventures including Texas Instruments/Acer are slowing their
spending.
We see this slowdown as a short-term situation. It may take six months for
the Asian financial dip to right itself and in the mean time, the slowdown
in production will allow demand to catch up with supply and the outlook will
improve.
Jeff Middleswart: The slowdown in SE Asia was not the trigger to problems in
the chip equipment industry. Rather, it was an accelerator as signs of a
slowdown were already being seen prior to the Asian crisis. In a survey
conducted by the Semiconductor Equipment Manufacturing Association that
addressed spending patterns over the past ten years, it was revealed that
for seven years, spending was under $10 billion per year, and in each of the
past three years, it has exceeded $40 billion. The consequence of this
dramatic rise in spending has been a huge, and precipitous increase in
capacity that has flooded the market with DRAM chips and has driven prices
markedly lower. As a result, supply is far ahead of demand. Add to the mix a
slowdown in Asia that will take away a growth market and a large mass of end
users, and fortunes look even more dim for the chip equipment companies
since orders for new products simply won't be there.
I don't think that there will be the necessary "boom" in exports to create
the need for more chip equipment or the construction of new plants. We
currently have a huge glut in virtually everything produced with
semiconductor capital equipment such as microprocessors, memory, disk
drives, modem chips, etc., and prices are plunging. There simply isn't
enough demand for PCs to absorb all this supply despite the much lower
prices. With Asia falling, and Asia was 25% of all PCs purchased worldwide
in the first half of 1997, I see end-demand falling further and there is
still more supply coming on-line, which will keep prices declining. As a
result, I don't know that any chip makers are going to be very profitable
for awhile regardless of how their currencies fall, and I don't see another
round of added capacity to be the answer. Moreover, even if companies in
Korea and Taiwan continue to buy as chip prices come down, I'm not confident
that they'll even have the money or the wherewithal to obtain bank financing
to buy more equipment. In sum, there will be no ensuing boom in exports to
necessitate a boom in imports of chip equipment. In fact, I would argue that
the chip equipment business is currently in a "bust" phase that could
potentially last the next two to three years.
Min Pang: There are two observations to be made with respect to the first
part of this question. First, and foremost, companies that have a business
exposure to the region will not suffer as much as people think. To
illustrate this, let's look at industry leader, Applied Materials, which
currently derives about 45% of its revenues from the Asia-Pacific region:
Japan (15%), Korea (10%), Taiwan (15%), and Singapore (5%). Taiwan and
Singapore are both economically strong, and more importantly, have been
among the least affected by the currency crisis that has severely impacted
several of their neighbors. Of these four countries, Korea holds the most
risk, but even if Korean businesses cut capital spending by 25%, that would
represent only a 2.5% drop in revenues for AMAT-- not a huge amount.
In actuality, chip equipment companies are suffering more from a perception
problem than from fundamental issues. With that said, the second point to
consider is stock price. Whether the company has limited exposure or
extensive exposure, stock prices in this group are suffering as a whole. I
would anticipate that trading in this area will be very volatile for a
quarter or two, but 6-9 months from now, investors will look back on this
episode and consider it just a blip in the grand scheme of things.
Given the pervasive uncertainty regarding Asian prospects, there could be
short-term spending pushouts and cancellations. We are in fact hearing about
these right now. Long-term, however, increased demand for equipment is
inevitable as the currency devaluation, and an ensuing boom in exports, will
create classic demand pull.
Briefing: Which semiconductor equipment companies have the greatest and
least exposure to SE Asia? Have you changed your earnings estimates or
ratings on any of the semiconductor equipment companies solely as a result
of recent events in SE Asia?
Jerry Fleming: Most semiconductor equipment companies have pretty wide
exposure in SE Asia, particularly in emerging markets like Taiwan and Korea.
Japan is the exception where exposure varies quite a bit by company. The
accepted model for the industry breaks down sales in the following regions:
US - 30%, Japan - 30%, all other SE Asia countries - 30%, and Europe - 10%.
That puts 60% of all equipment sales in the SE Asian region.
We have not yet changed our estimates because most of the companies we
follow are saying that they have not seen the rescheduling of shipments that
we are hearing about in the press. We have been fairly cautious as evidenced
by our downgrade in August when we took the group off our recommended list.
At that time, we saw the level of spending in Asia as an indication that
DRAM prices were going to fall and we felt that Korea and Taiwan were
potential sources for order cut backs. Typically, when equipment companies
spend more that 25% on plant and equipment, it is a warning sign. At the end
of this summer, spending in Korea and Taiwan was up around 40%. Spending in
Japan has been more conservative and spending in the US has been mixed.
Jeff Middleswart: The extent of exposure in SE Asia is hard to quantify, but
it is safe to say that all of the chip equipment companies will suffer
regardless of the level of exposure since the bulk of demand has come from
Asian plants. Yet, even if you take Asia out of the equation, this group was
subject for a fall because there exists a situation where more capacity in
the business is competing with a shrinking demand base.
I have not made any changes to earnings or ratings solely on account of SE
Asia since the business had already started to head lower prior to the Asian
debacle. Even when this industry was enjoying its bullish run during the
summer, I was still fairly bearish due to the capacity situation that was
simply overlooked by momentum investors, and the companies themselves as
many were sacrificing profits for market share. In addition, the slowdown in
the business was made apparent when Intel delayed its plans for construction
of its new plant, and when Micron Technology shut down construction of its
new plant in Utah in response to the dramatic decline in DRAM pricing.
Min Pang: On the surface, back-end companies such as Kulicke & Soffa and
Credence Systems appear to have the most exposure to SE Asia, but they
really don't. Conversely, front-end businesses, or wafer companies, seem to
have less exposure on a percentage basis, but actually have more in terms of
end use per cent. Take, for instance, Credence Systems which sells 65% of
its products to the SE Asia region. At first glance, this exposure is cause
for concern, yet it must be placed in its proper context. That can be done
by examining who its customers are. Credence sells to sub-contractors in the
region who test chips sent to them by semiconductor companies. Of the chips
tested, 95% are exported back to the U.S. and Europe. Thus the consideration
that has been overlooked is that these sub-contractors derive only 5% of
their revenues from the SE Asia region and Credence Systems' total end
exposure is 5% of 65%, or 3%. As such, they should not be largely impacted
by the currency crisis; moreover, the unit demand for chips remains high and
so the need for back end equipment hould also remain healthy.
Those companies with the least exposure to Asia are Photronics (about 5% and
then only to Singapore) and PRI Automation (10-15%). With respect to the
former, there is no reason on earth why its stock has gotten hit as hard as
it has since it has such limited exposure, and participates in a business
sector that is design-driven primarily rather than capacity-driven. Yet,
there now exists a negative bias for the group as a whole, and Photronics is
simply another domino that has fallen as a result of group sentiment.
I haven't changed any earnings estimates, but I have moved my rating on
Applied Materials from Strong Buy to Buy. The reason being is that there is
so much noise right now with respect to the troubles in the Asia-Pacific
region that it has become difficult for investors to separate psychology
from fundamentals. The downgrade, then, reflects a short-term view, yet AMAT
is still very attractive from a long-term perspective (3-4 quarters out).
Briefing: What is your expectation for the pace of orders growth in the
semiconductor equipment industry next year? What factors will contribute to
either a slow down or an increase in orders?
Jerry Fleming: There are two sets of variables impacting the growth rate for
this industry. The first is the currency/banking situation and DRAM prices
which we think will be short-term (next couple of quarters).
The more serious issue is that the group is on the verge of having to make
significant technological changes in the way chips are made. Wafer size is
going from 8 inches in diameter to 12 inches and the interconnect material
is going to copper from aluminum. This conversion will be expensive - one
and a half to two times the cost of prior chip making equipment but will
yield two and a half times as many chips per factory which will ultimately
be a boom. In the interim, the chip equipment companies will have to incur
the cost of developing and producing the new equipment and endure the lag
time while the chip makers first upgrade existing equipment and then slowly,
when the technology is perfected, make the capital expenditures for the
conversion.
We think that growth in the next two years will be relatively moderate. The
semiconductor equipment industry has been growing at roughly 17-18% for the
past 20 years. We are forecasting 5-10% growth in 1998 and 1999. We think
that the following segments will beat the industry average because they are
already selling equipment for the next generation and/or will benefit from
plant upgrades before the conversion.
(1) Lithography (imaging chip patterns on silicon): In this interim period,
chip makers will shrink chips so that they can get more chips per wafer and
to do so, they need smaller lenses. The large players are all foreign -
Nikon (NINOF), Canon (CANNY), and the Dutch company, ASM Lithography
(ASMLF). ASMLF does the best job in DeepUV, but it is number three in terms
of overall market share. The US way to play this part of the market is Cymer
(CYMI) which makes the laser that is the heart of each of these lithography
systems. The stock has not been cheap so we shied away from it. However,
since it has come down quite a bit (19 1/4 from its 52-week high of 49 1/4),
we may be adding it to our radar screen.
(2) Wafer inspection/ Metrology: The industry norm for inspection
historically was random sampling. In an effort to increase the number of
good chips per plant, companies are moving to increased inspection during
more of the steps in the manufacturing process. The solid number one player
in this arena is KLA-Tencor (KLAC). ADE Corp (ADEX) makes inspection
equipment for bare wafers and has as strong a sales position in that market
as KLAC has in sales to chip makers.
(3) Interconnect technology (deposition and etch equipment): Prior to the
transition to copper, chips will get smaller and the aluminum wires
connecting the transistors will get narrower but more of those wires will
need to be stacked on a chip to preserve speed. Novellus (NVLS), Applied
Materials (AMAT), and Lam Research (LRCX) are our favorites. Long-term,
Novellus wins as they have a good portfolio of products for the transition
phase and are one of the leaders in development of copper technology.
A new process has emerged in deposition called CMP (Chemical Mechanical
Planarazation) which will be used with copper as well. It smoothes and
polishes each layer which is necessary before another layer can be applied.
Integrated Process Equipment (IPEC) and SpeedFam International (SFAM) both
have this as their principal business. LRCX and AMAT have entered this
market, but IPEC and SFAM have the superior equipment. We may add SFAM to
our radar screen in the near future.
Jeff Middleswart: The pace of orders is going to slow tremendously given
that everyone is already screaming about too much capacity. While spending
on chip equipment had been running around $40 billion as of late, I would
anticipate that it will slide to around $15 billion in 1998 which represents
a 60% cutback. A significant portion of that total will come from the
completion of plants already under construction. After that, there will be a
scant number of plants being built to create the need for large orders of
chip equipment. Of course, changes in new technology will always require
existing plants to upgrade some equipment, so there will always be some
semiconductor equipment orders. However, to maintain the huge level of
orders this industry has enjoyed the last three years and investors have
come to accept as a given, chip companies have to keep building tons of new
plants. And at this time there is little economic justification to build a
bunch of new plants; and the biggest mavericks that build regardless of
economics, such as Korea and Taiwan, are now out of money.
Min Pang: The confusion surrounding the pace of orders is that many analysts
on the street had growth projections in excess of 20%, and in light of
recent developments, they have been forced to revise those estimates lower
to the 10-15% range which, by the way, is still pretty decent. Our corporate
spending projection for 1998 has been 11% since the beginning of 1997, and
we don't see anything at the moment to materially change that.
This pace of growth should be sustainable given the shift to 0.25 micron
manufacturing technology. Historically, the memory area accounts for 60% of
all capital spending. While the precipitous drop in DRAM prices has been
well-documented, rapid price erosion is a fact of life in this business, and
herein lies the Catch 22 for memory makers. Though prices are declining, and
they're not making money, it is incumbent on the memory companies to make
chips cheaper and to be at the forefront of design innovation. Thus they are
making chips smaller so that more can fit on a wafer. To achieve this end
they have to buy new technology, and with everyone wanting to be the first
to 0.25 micron, we will witness a lot of equipment being bought. Conversely,
sustained price pressure on DRAM chips could result in sporadic spending
patterns.
Briefing: Which companies are you recommending and/or avoiding?
Jerry Fleming: Please see above response.
Jeff Middleswart: I'm avoiding just about everything as these companies
usually lose money in a bust phase. In addition, most of the chip equipment
stocks have run ahead of fundamentals. Historically, these stocks trade at
book value or below so it is alarming to see Applied Materials trade at 4x
book, Novellus 4.5x book, Kulicke & Soffa at 2x book, and KLA-Tencor 3x
book. The latter, though overpriced, is the only component I would be
tempted to take a position in because this company makes equipment that
allows its customers, the chip companies, to improve their yields. Keep in
mind that this is a high fixed cost business, and when volume drops, profits
drop even quicker.
Min Pang: The answer to this question is best addressed in the following
way. If you're worried about Asia , then I would recommend Photronics (PLAB)
and PRI Automation (PRIA) which, as I said before, are relatively unexposed.
If you're worried about an actual downturn in the business, then I would
suggest Photronics (PLAB) and DuPont Photomasks (DPMI) since the mask
companies are more design-driven than capacity-driven, and that is why they
hold up well in business cycles. Finally, if you have a long-term outlook
(3-4 quarters), Applied Materials (AMAT) is an excellent candidate. I also
like Credence Systems (CMOS) because the back-end business is still pretty
healthy. In addition, Lam Research (LRCX) looks compelling at current
values. Of the chip equipment stocks that I follow, I have assigned the
following ratings: Photronics (Strong Buy); DuPont Photomasks (Strong Buy);
Credence Systems (Strong Buy); PRI Automation (Buy); Brooks Automation
(Buy); Kulicke & Soffa (Buy); Applied Materials (Buy); and Lam Research
(Buy). |
